Geospatial Health 2021; volume 16:904

Geo-epidemiological reporting and spatial clustering of the 10 most prevalent cancers in

Ebrahim Babaee,1 Gholamreza Roshandel,2,3 Meysam Olfatifar,4 Arash Tehrani-Banihashemi,1 Arezou Ashaari,1 Marzieh Nojomi1,5 1Department of Community Medicine, Preventive Medicine and Public Health Research Centre, Iran University of Medical Sciences, Tehran, Iran; 2Golestan Research Centre of Gastroenterology and Hepatology, Golestan University of Medical Sciences, Gorgan, Iran; 3Iranian National Population-Based Cancer Registry Secretariat, Cancer Office, Deputy of Health, Ministry of Health, Tehran, Iran; 4Gastroenterology and Liver Diseases Research Centre, Research Institute for Gastroenterology and Liver Diseases, Shahid Beheshti University of Medical Sciences, Tehran, Iran; 5Department of Sociology & Anthropology, Nipissing University, North Bay, ON, Canada

geographical underlying causes. The geographic distribution of Abstract cancers is reported as standardized incidence rates at the provin- Cancer is a problem of both global and local concern. We cial level considering risk with respect to sex and age. A geograph- determined the geo-epidemiological and spatial distribution of the ical information systems (GIS)only approach based on Anselin Local 10 most common cancers in Iran. We used the data of the Iranian Moran’s index method was used to map clusters and spatial auto- Cancer Registry for the year 2014 analysing the prevalence of correlation patterns. The mean age of the patients was 55.6 112,131 registered cancer cases with the aim of detecting potential (±17.8) and 61.7 (±18.2) for females and males, respectively, in the database whichuse showed 46.1% (n=51,665) of all cases to be female. Analysis of the spatial distribution of cancers showed sig- nificant differences among the different provinces. Stomach and Correspondence: Marzieh Nojomi, Preventive Medicine and Public breast cancers were the most prevalent cancers in men and Health Research Center, Psychosocial Health Research Institute, females, respectively. The highest incidence rates of stomach can- Community and Family Medicine Department, School of Medicine, cer were found in Ardabil and Zanjan provinces, with 48.38 and Iran University of Medical Sciences, Postal code: 1449614542, Tehran, 48.08 per 100,000 population, respectively, while Tehran and Iran. Fax: +98.2186703350. Yazd provinces had the highest incidences of breast cancer, 51.0 E-mail: [email protected] and 47.5 per 100,000 population, respectively. Strong clustering Key words: Cancer; epidemiology; spatial analysis; cluster analysis; patterns for stomach and breast cancers were identified in the Iran. north-western provinces and in , respectively. These patterns indicate a diversity of geo-epidemiological con- Acknowledgements and funding: this research was supported by the tributing factors to cancer incidence in Iran. Elite Researcher Grant Committee under award number of 982858 from the National Institute for Medical Research Development (NIMAD), Tehran, Iran. We would like to thank the Iranian national population- based cancer registry secretariat and Cancer Office of the Ministry of Introduction Health of Iran for their complete collaborations.Non-commercial Cancer is a universal public health burden lacking a common Conflict of interests: the corresponding author reports grants from the pre-emptive approach. The global burden of disease study in 2017 National Institutes for Medical Research Development (NIMAD), dur- (GBD 2017) reported 24.5 million incident cancer cases world- ing performing the study. All the other authors declare no conflict of wide (Global Burden of Disease Cancer Collaboration, 2019). interests. Bray et al. (2018) in their study refer to GLOBOCAN, which is another body providing estimates of cancer incidence, and report Received for publication: 7 June 2020. that the incidence rates of cancer and the related deaths in Asia are Revision received: 12 November 2020. 48.4% and 57.3%, respectively (The Lancet, 2018). According to Accepted for publication: 15 November 2020. previous studies, in developed countries, the prevalence of cancers ©Copyright: the Author(s), 2021 is predicted to reach 45% by the year 2025 (Kamangar et al., Licensee PAGEPress, Italy 2006). Notably, in low- and middle-income countries, population Geospatial Health 2021; 16:904 growth and aging are the major contributors to the increase of can- doi:10.4081/gh.2021.904 cer incidence (Global Burden of Disease Cancer Collaboration, 2019). Besides, more than 70% of all the cancer-related deaths This article is distributed under the terms of the Creative Commons occur in countries whose populations have a medium to the low- Attribution Noncommercial License (CC BY-NC 4.0) which permits any level standard of living as well as limited resources for the preven- noncommercial use, distribution, and reproduction in any medium, pro- vided the original author(s) and source are credited. tion, diagnosis, and treatment of cancer (Mousavi et al., 2008). In Iran, aging and increased related risk factors are known as

[page 80] [Geospatial Health 2021; 16:904] Article

the main causes of the increased incidence of cancer, which is the women, and out of 112,131 registered cases in total, we only third leading cause of death in Iran after coronary heart disease and excluded data from 216 patients due to incomplete information. accidents (Alireza et al., 2005; Mousavi et al., 2009). Given the The 3rd edition of the International Classification of Diseases importance of cancer control, effective strategies should be imple- for Oncology (ICD-O-3) was applied (Fritz, 2013) for tumour cod- mented by the health systems. Population-based cancer registries ing. Using demographic and geographical information, the play an important role in monitoring trends related to cancer inci- patients’ place of residence were specified and then recorded in the dence, prevalence and mortality over time among different areas dataset that showed the geographic distribution of the cancers and social groups. Although the data presented are accurate, it is a reported as age-standardized incidence rates (ASRs) at the provin- difficult method that needs a strong structure. In 2010, Iran estab- cial and sex and age subgroup levels. Based on the latest census lished the Iranian National Population-based Cancer Registry published in 2011 by Iran’s statistics centre, the population of Iran (INPCR) in order to collect and aggregate regional cancer data was considered as the reference population, with the world’s stan- (Roshandel et al., 2019). This system for cancer information col- dard population applied to estimate the ASRs. All provincial, age lection and surveillance is comprehensive, but it is equally impor- and sex specific incidence rates were expressed per 100,000 or tant that the data needed can be reported and retrieved in a practical 1,000,000 population (Segi, 1960). way. Cancer maps could help to determine the regions with the Geospatial analysis highest density of cancer cases and also facilitate the investigation For geo-reporting we used the geographic map of Iran in of environmental risk factors of specific cancer types in different shape-file format. This file was obtained from the national carto- regions according to the Guide for Effective Programs, publicised graphic centre of Iran (https://www.ncc.ir/en-us/NCCPortal/1/ by the World Health Organization (WHO) (https://www.who.int/ page/HomePage). The geographic reports of cancer cases were cancer/ modules/en/). Geographical information systems (GIS) can based on three sections. Firstly, we generally symbolized the geo- perform these goals well (Scholten et al., 1991) through applica- demographical distributionsonly of the TMPCs in separate maps in tions that have since developed in different directions (Bergquist & terms of the participants’ gender using quantities expressed by Manada, 2019). gradually shaded colours. The incidence rates were reported in 5 Notably, creating geo-referenced maps is necessary to obtain categories on the maps based on quantities based the natural breaks valuable information about cancer cases concerning frequency of (Jenks) classificationuse method. In order to map clusters and spatial spatial features (Yomralioglu et al., 2009). In this regard, spatial autocorrelation patterns, we applied the Anselin Local Moran’s analyses help to generate new hypotheses. However, it is also index method. For the conceptualization of the spatial relationships important that clusters associated with certain cumulative risks we selected the fixed distance band option, which imposes a sphere such as income, urbanization and dietary patterns or areas that of influence where each feature is analysed within the context of require the targeted interventions can be found (Bafandeh & the neighbouring features located within the distance specified by Farhang, 2009; Hao et al., 2011). the threshold distance (distance band). For analysing the obtained So far, few studies have been conducted on this subject in Iran data, we used Stata software version 14 (Stata Corp, College and those available were carried out at the provincial level and Station, TX, USA) and Microsoft Excel Worksheet. The ArcGIS only on some types of cancer (Chamanparaa et al., 2015). In this software, v.10.2.2 (ESRI, Redlands, CA, USA) was applied to study, we report the results of the demographical and geographical combine the data and produce the spatial analysis. All these analy- distributions of the 10 most prevalent cancers (TMPC) in Iranian ses were completed in the ArcMap environment. men and women. We also investigated spatial clustering and auto- correlation patterns of the TMPCs (with regard to both genders) among the to answer the question whether neigh- bouring provinces tend to have similar incidence rates and whether Results they also form clusters based on such correlations. Non-commercialThe mean age of all patients (female and male) were 55.6 (±17.8) and 61.7 (±18.2) years, respectively, and the female share of the cancer cases investigated were 46.1% (n=51,665). Those Materials and methods being 75 years old or older constituted the largest age group (n=24,446, 21.8%), while the youngest group (≤4 years) only included 952 of the cases (0.8%). Cancer of the stomach (n=7149) Study area and breast (n=13,120) were the most prevalent cancers among This geographical registry-based study was conducted in 2020 Iranian men and women, respectively. The order of all the most in Iran. Located in the western part of Asia bordering the Caspian common cancers for each sex is reported in Table 1. In Tables 2 Sea, Persian Gulf and the Gulf of Oman, the country consists of 31 and 3, we report the ASRs as well as the sex-standardized inci- provinces (Figure 1) that have shown a great variability in cancer dence rates (per 1,000,000) of the TMPCs in men and women, incidence, in particular with reference to the type of cancer. respectively. The male incidence rate of stomach cancer was 1.1 and 2393.4 in the age groups 0-4 and ≥75 years, respectively, while the female incidence rate of breast cancer was 0.2 and 758.9 in the Data source and rates same age groups. Notably, the incidence of all cancers in both We used data for 2014 from INPCR that were released in 2019 sexes increased with age, except for uterus, breast and ovarian can- (Roshandel et al., 2019). Our study covered only 30 of Iran’s 31 cer. In these three cancers, the highest incidence rates were found provinces since information from was unavailable. to be between the ages of 50 and 70 years, with a decreasing trend Still, the data covered 98% of the Iran’s population, men and beginning from the age of 70 years. In men, cancer of the prostate

[Geospatial Health 2021; 16:904] [page 81] Article

amounted to 10.1% (n=5946) and colorectal cancer to 9.6% (Figure 5). The results of Local Moran’s I for all the different can- (n=5648), while in women colorectal cancer was 8.6% (n=4220) cers among men and women are shown in Figures 4 and 5. and stomach cancer 6.8% (n=3348); these were the second and third most prevalent cancers overall, respectively.

Geographical distribution and autocorrelation The geographical distribution of the standardized incidence Table 1. The most common cancers in Iranian males and females rates of TMPCs for men and women at the provincial level are 2014. shown in Figures 2 and 3, respectively. Based on these results, Type of cancer Number Percent Ardabil (48.38 per 100,000) and Zanjan (48.08 per 100,000) were the provinces with the highest incidence rates of stomach cancer in Male Stomach 7149 12.10 men occupying the first and the second rank, respectively. Among Prostate 5946 10.06 women, Tehran (51.0 per 100,000) and Yazd (47.5 per 100,000) Colorectal 5648 9.56 were the provinces with the highest incidence rates of breast can- Bladder 4825 8.16 cer. Additionally, colorectal and stomach cancers were the second Lung 4217 7.14 Leukaemia 2876 4.87 and third most prevalent cancers in Iranian women. Moreover, col- Brain, nervous system 2392 4.05 orectal cancer was mainly prevalent in the provinces Semnan and Oesophagus 2177 3.68 Tehran. Notably, Ardabil, Zanjan, East Azerbaijan and West Non-Hodgkin* 1981 3.35 Azerbaijan had the highest incidence rates of stomach cancer. Larynx 1417 2.40 When we turned our attention to the spatial cluster patterns of Other sites 20,477 34.64 all cancers under study for both genders, the high and low cluster Total 59,102 100.00 points produced by the Moran’s I showed different distribution pat- Female Breast only 13,120 26.75 terns depending on cancer type in the different provinces. For men, Colorectal 4220 8.60 stomach cancer was significantly clustered with high-high cluster- Stomach 3348 6.83 ing patterns in West Azerbaijan, East Azerbaijan, Ardabil and Thyroid 2911 5.94 Zanjan provinces (Figure 4). However, we observed no such pattern Leukaemia 1879 3.83 for prostate cancer, but high-high clusters of larynx cancer were Lunguse 1839 3.75 seen in North Khorasan and Razavi Khorasan in north- Brain, nervous system 1800 3.67 and a low-low pattern in Hamadan and Lorestan in the western part Oesophagus 1778 3.63 of the country (Figure 4). With regard to women, we observed low- Uterus 1726 3.52 low clusters of breast cancer in Ardabil and Kurdistan and high- Ovary 1653 3.37 Other sites 14,773 30.12 high ones in Semnan, Isfahan and Yazd, which are Iran’s central Total 49,047 100.00 provinces (Figure 5). For uterus cancer, we identified a low-low *Non-Hodgkin lymphoma. cluster pattern in and an outlier in

Table 2. Standardized age-sex specific incidence rates (per 1,000,000) of the most common cancers in Iranian males 2014. Age group Stomach Prostate Colorectal Bladder Lung Leukaemia Nervous Oesophagus Non-Hodgkin Larynx system* lymphoma 0-4 1.1 0.3 0.3 0.3 0.8 59.2 19.5 0 6.8 0 5-9 0.6 0.3 0.3 1.2 1.2 62.7 20.4 0.3 11.9 0 10-14 0.7 0 Non-commercial 0.7 0.7 1.4 39.2 14.8 0 11.7 0 15-19 1.1 0.7 5.4 3.9 3.6 46.3 19.4 0 16.2 0.4 20-24 7.7 1.5 8.7 4.9 7.1 36.8 27.5 0.9 22.2 0.3 25-29 13,0 1.7 19.8 7.2 5.8 27 30.9 2.4 24.6 1.0 30-34 18.2 1.4 32.7 13.6 13.1 24.2 35.7 4.1 23.5 2.1 35-39 33.3 2,0 58.5 22.4 14.3 31.9 45.6 10.1 31.6 4.2 40-44 72.1 7.8 89.9 46.2 42.3 41.6 60.8 19.5 45.5 17.1 45-49 124.3 20.4 145.0 92.9 83.5 56.6 70.1 37.9 55.0 38.7 50-54 259.7 109.8 272.3 212.6 183.2 86.1 105.3 70.4 90.1 85.0 55-59 438.3 291.6 415.6 358.1 285.0 129.9 143.7 122.2 120.4 135.9 60-64 642.6 629.1 604.3 539.7 459.8 174.8 182.0 196.4 154.1 182 65-69 1128.7 1131.2 743.0 728.1 666.3 241.1 223.8 336.3 185.4 225 70-74 1630.2 1735.2 1132.9 1003.3 882.5 302.9 206.6 523.5 231.1 238.1 ≥75 2393.4 2383.7 1285.0 1370.7 1225.5 514.8 294.4 778 323.7 294.4 Total number° 7145 5944 5644 4823 4214 2873 2390 2147 1981 1416 *Including the brain; °the number may be less than total due to missing data.

[page 82] [Geospatial Health 2021; 16:904] Article

have similar incidence rates and formation clusters based on this Discussion correlation, we also investigated the spatial clustering and autocor- In this study, we reported the geo-epidemiological distribution relation patterns of TMPCs (for both sexes) among the provinces of the TMPCs among Iranian men and women using population- of Iran. based cancer registries data. Iranian provinces were ranked based In doing so, we found some controversial incidences for cancer on the incidence of all types of cancers and the results were visu- of the ovary, breast and prostate in children less than 4 years old, ally displayed on a map. Moreover, we showed geo-epidemiologi- which should be interpreted with caution. It was further shown that cal distribution, spatial analysis of the incidence of TMPCs and stomach cancer is the most prevalent cancer among Iranian men, clustered patterns on a provincial level using national data in Iran. with the highest incidences in Ardabil and Zanjan provinces. In order to determine the tendency of neighbouring provinces to Previous studies have reported intake of high salt and opium

only use

Figure 1. The geographical situation of Iran.

Table 3. Age-sex specific incidence rates (per 1,000,000) of the most common cancers in Iranian females 2014. Age group Breast Colorectal Stomach Thyroid Leukaemia Lung Nervous system* Oesophagus Uterus Ovary 0-4 0.2 0.2 0 0 43 0.9 16.3 0 0 0.3 5-9 0 0.3 0.3 0.6 52.5 0.3 20.2 0 0 1.6 10-14 0.4 0.7Non-commercial 0 6.5 27.0 1.4 14.4 0 0 1.8 15-19 2.2 3 3 18.3 33.7 2.6 12.7 2.2 0.4 8.2 20-24 18.1 8.2 5.7 63.1 26.9 4.1 19.3 0.6 2.2 11.4 25-29 64.6 17.7 9.1 83.3 22.7 5.7 25.6 1 6.9 17.3 30-34 185.0 24.4 16.2 89.9 19.0 10.8 33.3 5.2 12.2 21.1 35-39 370.5 47.0 30 101.9 23.1 8.4 29.4 11.5 26.5 28.6 40-44 658.2 88.0 48.8 116.5 36.6 26.6 49.2 17 51.4 48.1 45-49 887.0 133.7 71.9 122.3 35.7 43.7 56.3 34.9 74.8 82.4 50-54 999.3 252.8 110.7 138.4 59.0 74.3 84.6 72.8 115.4 114.3 55-59 935.3 324.3 186.8 110.1 85.1 111.9 85.1 110.1 177.3 113.1 60-64 951.2 437.2 317.1 122.5 119.4 173.7 108.5 179.1 210.1 144.2 65-69 920.7 503.0 457.6 138.5 131.8 236.0 140.7 264.8 197.2 174.0 70-74 859.0 672.7 662.8 143.4 168.2 364.4 194.6 415.5 166.5 158.3 ≥75 758.9 854.0 1096.5 143.2 256.1 570.7 245.6 545.6 162 150.5 Total number° 13,116 4215 3343 2910 1879 1839 1779 1774 1725 1650 *Including the brain; °the number may be less than total due to missing data.

[Geospatial Health 2021; 16:904] [page 83] Article

only use

Non-commercial

Figure 2. The geographic distribution in 2014 of the age-standardized incidence rates (ASR) per 100,000 population of the most com- mon cancers in Iranian men.

[page 84] [Geospatial Health 2021; 16:904] Article

only use

Non-commercial

Figure 3. The geographic distribution in 2014 of the age-standardized incidence rates (ASR) per 100,000 population of the most com- mon cancers in Iranian women.

[Geospatial Health 2021; 16:904] [page 85] Article

only use

Non-commercial

Figure 4. Spatial cluster patterns of the most common cancers in Iranian males in 2014 based on Moran’s I.

[page 86] [Geospatial Health 2021; 16:904] Article

only use

Non-commercial

Figure 5. Spatial cluster patterns of the most common cancers in Iranian females in 2014 based on Moran’s I.

[Geospatial Health 2021; 16:904] [page 87] Article

intake, as well as helicobacter infection, as risk factors of stomach environmental factors as well as geographic and climatological cancer in these regions (Abdi et al., 2016; Yusefi et al., 2018). agents on cancer incidence (Yomralioglu et al., 2009). As men- Several central provinces and most southern provinces had low tioned earlier, few studies have been conducted in this regard in incidence rates of this cancer. Notably, the highest incidence rates Iran so far, and most of them were performed at the provincial were in men and women aged 70 years and older. level and on a small number of cancer types (Chamanparaa et al., Prostate and colorectal cancers were the second and third most 2015). After performing the analysis, spatial autocorrelation by prevalent cancers among Iranian men. It is noteworthy that the Moran’s I showed significant clustering patterns in different highest incidence rates of prostate cancer were in Yazd, Fars, and regions for different cancers. In men, Moran’s I showed stomach Tehran provinces. This cancer is particularly prevalent in devel- cancer as significantly clustered with high-high spatial patterns in oped countries (Baade et al., 2009), and a meta-analysis of 24 stud- West Azerbaijan, East Azerbaijan, Ardabil and Zanjan provinces. ies indicates that smoking is the most important risk factor This indicates that the spatial distribution of this type of cancer (Huncharek et al., 2010). The incidence of prostate cancer has also tends to cluster across the north-western provinces of Iran. In other been shown to be associated with the heavy metals such as copper, words, it seems that this clustering could be caused by common iron, and manganese (Karimi et al., 2012). With respect to colorec- contributing risk factors across among these provinces. tal cancer, we found its highest prevalence to be among Iranian For prostate cancer, there was no clustering pattern in any men in Semnan, Tehran and Yazd provinces, respectively. A study province. In 2006, a study performed in the UK showed a geo- was conducted in Khorasan-Razavi province in 2019 to identify graphic variation in prostate cancer incidence across different the potential contributing factors of colorectal cancer. Accordingly, counties, a variance possibly due to social, political, economic, this study the incidence of colorectal cancer is significantly associ- health care and environmental differences (Gregorio et al., 2006). ated with a high obesity and low fibre intake level (Goshayeshi et We also identified a high-high clustering pattern for colorectal can- al., 2019). Future studies along these lines will be very valuable to cer in men close to the centre of Iran in Semnan and Yazd determine the environmental (Jacquez et al., 2003) and hereditary provinces. Between 2007 andonly 2014, a study was conducted using (Goshayeshi et al., 2017) contributing factors of colorectal cancer Anselin Local Moran’s I to identify clusters and outliers of col- in the hotspot areas. orectal cancer distribution in (Halimi et al., Larynx cancer is one of the least prevalent cancers among men, 2020). They demonstrated significant geographical disparities in consequently ranked in the tenth place in our study. This cancer is colorectal cancer incidence with three significant clusters of both mostly prevalent in North . Based on the results high and low incidenceuse rates in Hamadan Province. Another study of a study performed in 2016 in Iran, larynx cancer was consider- performed in Khorasan-Razavi Province identified three signifi- ably common in men, and the estimated age-standardized inci- cantly high-low outlier areas as well as four significant low-high dence rates were 21.6 and 3.2 per 100,000 in men and women, outliers (Goshayeshi et al., 2019). respectively (Mirzaei et al., 2016). In addition, a systematic review In our study, lung cancer significantly clustered with a high- study in Iran confirms that larynx cancer is more common among high pattern in the Northwest of the country and a low-high outlier men than women (Hassanipour et al., 2019). pattern for men in the Southeast. This pattern of outliers indicates In Iranian women, breast cancer is known as the most preva- that low-value areas are overlapped with high-value ones. lent cancer (Roshandel et al., 2019), and women in Iran are affect- Furthermore, we identified a high-low outlier region of leukaemia ed at least a decade earlier compared to women in developed coun- occurrence in . A study performed in 2007 in tries (Nematolahi & Ayatollahi, 2017). Tehran, Yazd, Semnan, Tehran showed some evidence of clustering of leukaemia in chil- Khorasan-Razavi, and Isfahan provinces were found to have the dren in the central region of Tehran (Mosavi-Jarrahi et al., 2007). highest incidence rates of breast cancer, in particular affecting They showed that environmental factors may be strong risk factors women in the age group of 50-54 years and older. Other recent for leukaemia. studies in Iran in 2020, though with a smaller sample size than our In terms of breast cancer in women, we observed a low-low study, report the average age of women with breast cancer to be clustering pattern in Ardabil and Kurdistan provinces as well as a 54.4 years old (Dolatkhah et al., 2020).Non-commercial high-high one in the centre of Iran. Besides the known risk factors Although Roshandel et al. (2019) report ovarian cancer as the of breast cancer, previous studies have proven that cadmium, as a 9th most prevalent cancer before uterus cancer among Iranian metalloestrogen, can be considered an environmental risk factor women, we found this cancer to be the 10th most prevalent cancer for breast cancer (Byrne et al., 2013). Similar to men, stomach can- after uterus cancer in Iranian women. This can be due to subdivi- cer in women was significantly clustered with high-high patterns sion of the codes used for the different components of the female in north-western provinces. Moreover, female lung and ovary can- genital organs, which might have caused a difference in the num- cers only clustered as a high-low outlier in Bushehr and East ber of cancers of the internal organs. It seems therefore that the Azerbaijan provinces, respectively. incidence rates of some cancers in some provinces should be inter- We declare that several potential methodological limitations preted with caution. Indeed, some southern and south-eastern affected our study. This research was a registry-based study and provinces showed unusually low incidence rates of most cancers, this type of study have certain limitations in terms of the accuracy which may be due to underreporting detected cancers by south- of data recording (Hailer, 2015). Another limitation was the use of eastern provinces or reporting of the cases of these provinces by the census population data (with a 5-year interval, instead of one- other provinces. Because most patients are referred to other year intervals) as dominator, to calculate the ASRs. The census is provinces due to the lack of diagnostic facilities in the south-east- conducted in Iran every 5 years. If the census is conducted at one- ern provinces. year intervals, more accurate demographic information may be In this study, we used exploratory spatial data analysis to deter- available for health calculations. Incorrect recording of residence mine spatial distribution patterns of cancers. The findings of this addresses of patients was another possible limitation of this study, analysis can be helpful when investigating the potential effects of in which the patient’s place of diagnosis may be recorded as the

[page 88] [Geospatial Health 2021; 16:904] Article

patient’s place of residence. Goshayeshi L, Khooiee A, Ghaffarzadegan K, Khorram MR, It is noteworthy that when recording patient information, Bishehsari F, Hoseini B, Rezayat KA, Esmaeilzadeh A, many organizations and medical institutions do not record com- Mozaffari HM, Ghanayee O, 2017. Screening for Lynch syn- plete geographic data. Correspondingly, this can be considered as drome in cases with colorectal carcinoma from Mashhad. AIM an important limitation when using GIS and spatial data for public 20:332-7. health settings. Goshayeshi L, Pourahmadi A, Ghayour-Mobarhan M, Hashtarkhani S, Karimian S, Dastjerdi R, Eghbali B, Seyfi E, Kiani B, 2019. Colorectal cancer risk factors in north-eastern Iran: A retrospective cross-sectional study based on geograph- Conclusions ical information systems, spatial autocorrelation and regres- sion analysis. Geospat Health 14:793. Complete and usable significant spatial and autocorrelation pat- Gregorio DI, Samociuk H, DeChello L, Swede H, 2006. Effects of terns of cancer incidence in Iran have been demonstrated. Analysis study area size on geographic characterizations of health of the spatial distribution of cancers showed significant differences events: prostate cancer incidence in Southern New England, among different provinces. To investigate the environmental risk USA, 1994-1998. Int J Health Geograph 5:8. factors of cancers, researchers can consider the findings of this study Hailer NP, 2015. Orthopedic registry research-limitations and in their future epidemiological studies. The findings of our study can future perspectives. Acta Orthop 86:1. also provide useful information for policymakers to plan appropriate Halimi L, Bagheri N, Hoseini B, Hashtarkhani S, Goshayeshi L, screening programs and cancer surveillance. Kiani B, 2020. Spatial analysis of colorectal cancer incidence in Hamadan Province, Iran: a retrospective cross-sectional study. Appl Spat Analy Policy 13:293-303. Hao Y, Landrine H, Jemalonly A, Ward KC, Bayakly AR, Young JL, References Flanders WD, Ward EM, 2011. Race, neighbourhood charac- Abdi E, Latifi-Navid S, Yazdanbod A, Zahri S, 2016. Helicobacter teristics and disparities in chemotherapy for colorectal cancer. pylori babA2 positivity predicts risk of gastric cancer in JECH 65:211-17. Ardabil, a very high-risk area in Iran. Asian Pac J Cancer Prev Hassanipour S, Delam H, Nikbakht HA, Abdzadeh E, Salehiniya 17:733-8. H, Arab-Zozaniuse M, Ghaem H, 2019. The incidence of laryn- Alireza S, Mehdi N, Ali M, Alireza M, Reza M, Parkin D, 2005. geal cancer in Iran: a systematic review and meta-analysis. Cancer occurrence in Iran in 2002, an international perspec- CEGH 7:457-63. tive. Asian Pac J Cancer Prev 6:359. Huncharek M, Haddock KS, Reid R, Kupelnick B, 2010. Smoking Baade PD, Youlden DR, Krnjacki LJ, 2009. International epidemi- as a risk factor for prostate cancer: a meta-analysis of 24 ology of prostate cancer: Geographical distribution and secular prospective cohort studies. AJPH 100:693-701. trends. Mol Nutr Food Res 53:171-84. Jacquez GM, Greiling DA, 2003. Geographic boundaries in breast, Bafandeh Y, Farhang S, 2009. Subsite distribution of gastric cancer lung and colorectal cancers in relation to exposure to air toxics in an area of high prevalence-northwest Iran. J Epidemiol in Long Island, New York. Int J Health Geograph 2:4. 19:202-5. Kamangar F, Dores GM, Anderson WF, 2006. Patterns of cancer Bergquist R, Manada S, 2019. The world in your hands: GeoHealth incidence, mortality, and prevalence across five continents: then and now. Geospat Health 14:779. defining priorities to reduce cancer disparities in different geo- Bray F, Ferlay J, Soerjomataram I, Rebecca L, Siegel RL, Torre graphic regions of the world. J Clin Oncol 24:2137-50. LA, Jemal A, 2018. Global cancer statistics 2018: GLOBO- Karimi G, Shahar S, Homayouni N, Rajikan R, Bakar NF, Othman CAN estimates of incidence and mortality worldwide for 36 MS, 2012. Association between trace element and heavy metal cancers in 185 countries. Ca Cancer J Clin 68:394-424. levels in hair and nail with prostate cancer. APJCP 13:4249-53. Byrne C, Divekar SD, Storchan Non-commercialGB, Parodi DA, Martin MB, 2013. Mirzaei M, Hosseini SA, Ghoncheh M, Soheilipour M, Soltani S, Metals and breast cancer. J Mammary Gland Biol Neoplasia Soheilipour F, Salehiniya H, 2016. Epidemiology and trend of 18:63-73. head and neck cancers in Iran. Global J Health Sci 8:189. Chamanparaa P, Moghimbeigi A, Faradmal J, Poorolajal J, 2015. Mosavi-Jarrahi A, Moini M, Mohagheghi MA, Alebouyeh M, Exploring the spatial patterns of three prevalent cancer latent Yazdizadeh B, Shahabian A, Nahvijo A, Alizadeh R, 2007. risk factors in Iran; using a shared component model. IJER Clustering of childhood cancer in the inner city of Tehran 2:68-77. metropolitan area: a GIS-based analysis. Int J Hyg Environ Global Burden of Disease Cancer Collaboration, 2019. Global, Health 210:113-9. regional, and national cancer incidence, mortality, years of Mousavi SM, Davanlou M, Gouya MM, Hajsadeghi N, Ramazani life lost, years lived with disability, and disability-adjusted R, Seddighi Z, 2008. Cancer incidence and mortality in Iran. life-years for 29 cancer groups, 1990 to 2017: a systematic Ann Oncol 20:556-63. analysis for the Global Burden of Disease Study. JAMA Mousavi SM, Gouya MM, Ramazani R, Davanlou M, Hajsadeghi Oncol 5:1749-68. N, Seddighi Z, 2009. Cancer incidence and mortality in Iran. Dolatkhah R, Somi MH, Jafarabadi MA, Hosseinalifam M, Sepahi Ann Oncol 20:556-63. S, Belalzadeh M, Nezamdoust M, Dastgiri S, 2020. Breast can- Nematolahi S, Ayatollahi MT, 2017. A comparison of breast cancer cer survival and incidence: 10 years cancer registry data in the survival among young, middle-aged, and elderly patients in northwest. Int J Breast Cancer 2020:1963814. southern Iran using Cox and empirical Bayesian additive haz- Fritz AG, 2013. International classification of diseases for oncolo- ard models. Epidemiol Health 39. gy: ICD-O. World Health Organization, Geneva, Switzerland. Roshandel G, Ghanbari-Motlagh A, Partovipour E, Salavati F,

[Geospatial Health 2021; 16:904] [page 89] Article

Hasanpour-Heidari S, Mohammadi G, Khoshaabi M, Sadjadi The Lancet, 2018. GLOBOCAN 2018: counting the toll of cancer. A, Davanlou M, Tavangar SM, 2019. Cancer incidence in Iran Lancet 392:985. in 2014: results of the Iranian National Population-based Yomralioglu T, Colak EH, Aydinoglu AC, 2009. Geo-relationship Cancer Registry. Cancer Epidemiol 61:50-8. between cancer cases and the environment by GIS: A case Scholten H, de Lepper MJ, 1991. The benefits of the application of study of Trabzon in Turkey. Int J Environ Res Public Health geographical information systems in public and environmental health. World Health Stat Q 44:160-70. 6:3190-204. Segi M, 1960. Cancer mortality for selected sites in 24 countries Yusefi A, Lankarani KB, Bastani P, Radinmanesh M, Kavosi Z, (1950-57). Department of Public Health, Tohoku University of 2018. Risk factors for gastric cancer: a systematic review. Medicine, Sendai, Japan. APJCP 19:591.

only use

Non-commercial

[page 90] [Geospatial Health 2021; 16:904]